3

As this is my first post, I hope that what I write is pretty clear and isn't too disappointing. Without further ado, I am quite curious with regards to my understanding of Disjoint Union Spaces. (I am currently self studying from Introduction to Topological Manifolds by Lee.)

First, I will begin with a definition.

Def: Suppose $(X_a)_{a\in A}$ is an indexed family of non-empty topological spaces. Denote their disjoint union by $\coprod_{a\in A}X_a$.

We define the $\textbf{canonical injection}$ by: $\iota_a: X_a\rightarrow \coprod_aX_a$ with the assignment being $\iota_a(x)=(x,a)$.

As a matter of convention, we force the equality: $X_a= \iota_a(X_a)$.

We define define a topology on $\coprod_aX_a$ by declaring that $U\subseteq \coprod_aX_a$ is open if and only if $U\cap X_a$ is open in $X_a$ for each a.

...................

Lee, in his book, states that $X_a$ in $U\cap X_a$ is considered as a subset of the disjoint union. Does this mean, ignoring the equality convention above, that he really means $\iota_a^{-1}(U)$ is open in $X_a$, for each a?

Secondly, what is the purpose of the convention above?. Lastly, what are good ways to start thinking about the disjoint union?

  • 1
    A way to think about disjoint unions: you can think of the disjoint union of (for example) a sphere of radius 1 and a sphere of radius 2 as: the set formed by the sphere of radius 1 centered at (0,0,0) and the sphere of radius 2 centered at (10,10,10): the two are far apart from each other, open sets in the union are just formed by taking the union of one open set in each one, and the same for closed sets. Is that the sort of answer you're looking for? – John Palmieri May 13 '20 at 22:21

1 Answers1

1

In general, let $X$ be a set and $(X_i, \tau_i)_{i\in I}$ a family of topological spaces. Every family $\{f_i: X_i\longrightarrow X\}_{i\in I}$ of mappings induce a topology on $X$ called the final topology with respect to this family.

By definition, the final topology on $X$ with respect to the family $\{f_i: X_i\longrightarrow X\}_{i\in I}$ is the finest topology on $X$ which turns each $f_i: X_i\longrightarrow X$ into a continuous map.

It turns out, $U$ is open in this topology if and only if $f_i^{-1}(U)$ is open in $X_i$ for every $i\in I$.

The disjoint union topology is a special case, can you see why?

See final topology for further details.

You may also want to check the dual notion called initial topology.

As to the convention, notice

$$\imath_a: X_a\longrightarrow \imath_a(X_a)$$ are bijections. In those cases, it is usual to identity the domain with its image to simplify the notations.

Finally, you should think of the disjoint union as a way to force the members of the union to be disjoint, since:

$$(i, X_i)=(j, X_j)\Leftrightarrow i=j.$$ Think of every element of the indexing set $A$ as a point in the plane and sitting on the point $i\in A$ there is the set $X_i$.

PtF
  • 9,655
  • Hi, I believe this only answers one of my questions. Could you perhaps answer the other 2? – Moe Khaled Bin-Lateef May 13 '20 at 21:33
  • As to the convention, the mappings $\imath_a$ are injective, hence are bijections $\imath_a:X_a\longrightarrow \imath_a(X_a)$. In those cases, it is usual to identify the domain with its image. Explicitly, $\imath_a(x):=(a, x)$. – PtF May 13 '20 at 21:35